APRSISCE Lock Fix & New Features!
Lynn Deffenbaugh
Group,
There's another new version of APRSISCE for Windows Mobile just posted to the Files section. I discovered a fairly serious resource consumption that causes Windows Mobile to run out of resources and lock up if you drive around at a low zoom level causing lots of OpenStreetMap files to paint. This resource leak has been fixed along with the following changes. 1) Reworked the APRS-IS network handling to improve responsiveness and reduce load on the device. 2) Double-clicking the "APRS OK" window no longer brings up a series of network devices, but instead shows traffic statistics for the APRS-IS connection including packets and bytes sent and received for this session. 3) Reworked where stations are painted on the OSM tiles. They weren't quite landing on streets at larger zoom levels. It uses the old way until all tiles have been filled in, so don't be alarmed if you see the stations "jump" right after the map completes. You should only see this the first time you're in an area while the OSM tiles are loading from the network instead of from the cache. 4) The OSM map is now clipped to the old circle and doesn't overflow around and behind the various other windows. 5) Double-clicking an empty space on the map will popup OpenStreetMap tile statistics if any tiles have been loaded in that session. Eventually, I hope to also show the status of the on-device cache of tiles by zoom level. 6) Increased the available zoom levels. On my 320x240 AT&T Tilt, I can now zoom out to 524 miles instead of the old 256. 7) Reworked how tiles are cached on CE. You might want to locate the OSMTiles folder (probably in the root of your device) and delete it and all files within it. The new caching layer uses directories that exactly mirror OSM's URLs. 8) To store the OSM cache on a Storage Card (assuming you have one), close APRSISCE and edit the APRSISCE.xml file that should have synced to your desktop. Locate the <OSM.Path> argument and set it to <OSM.Path>/Storage Card/OSMTiles/</OSM.Path>. Save the file and wait for it to sync back to your device before running APRSISCE. You should then see an OSMTiles folder on your Storage Card. 9) The OSM Tile Cache now prefetches tiles in anticipation of your movement or zooming. It fetchs one tile out in each ordinal direction from the screen along with all tiles one closer zoom below and one further zoom above. 10) While APRSISCE is fetching OSM tiles, you'll see a yellow circle expanding from the center of the map. Each jump is one tile. When it reaches the outer circle, it has worked off all queued prefetches. It prioritizes the queue, so it'll still get the missing visible tiles first. Let me know what you think and let me know what you'd like to see enhanced and/or changed, especially if you encounter any bugs! Lynn (D) - Author of APRSISCE - (APRS message me at KJ4ERJ-12) |
|
John Joseph <josephj@...>
Looks good Lynn!
The map detail is a little hard to read. It is light gray on a white background on my touch pro. I don’t know whether there is a setting that I can change. The GPS sats on the bottom are still scrunched. It would be nice to have an option to not show the GPS status. The velocity and direction of a selected station would be nice as well.
N8JJ
From:
aprsisce@... [mailto:aprsisce@...] On Behalf Of ldeffenb
Group,
No virus
found in this incoming message. |
|
Lynn Deffenbaugh
John,
toggle quoted message
Show quoted text
The left/right arrows will change the transparency of the OSM maps. If that doesn't do it, the <OSM.Percent> is a number from 0 (no maps) to 100 (full quality maps) in the XML file. I need to work on the screen layout in general. Actually, though, for the station overlay characters, I'm doing different things with fonts, so I might be able to at least make the satellite number fit the box. I'll think about making those optional, but I know that I find them useful when I'm mobile as it let's me know if I've lost coverage or am about to lose coverage. Unfortunately, that's an all to common event with the Tilt's GPS. Remember that not all stations include velocity and direction in their beacon. APRS.FI (for instance) attempts to calculate it with delta distance over delta time, but it all to often comes up with some creative numbers (like cruising the Interstate at 120+ mph). What I'm hoping to work out for the next release is to allow you to select a station and make it the center, moving/scrolling the maps on each beacon. I think that would be kinda handy when I'm sitting somewhere waiting for someone to join me! Thanks for the feedback and let me know if anything goes worng when you take it mobile! Lynn (D) - KJ4ERJ John Joseph wrote:
|
|
Noel Shrum
I'd like to see larger icons on VGA devices.
Its kinda hard to tell what the icons are when you have a high DPI screen. Noel |
|
David Shrum
Lynn,
toggle quoted message
Show quoted text
I downloaded it and did the path changes you suggested; I just copied them and pasted them into the 'xml' file. I ran it from home just so it would load what it needed. I have a good size card so I doubt I'll have any problems there, thanks. Dave N8PU -----Original Message-----
From: aprsisce@... [mailto:aprsisce@...] On Behalf Of ldeffenb Sent: Sunday, August 09, 2009 8:45 PM To: aprsisce@... Subject: [aprsisce] APRSISCE Lock Fix & New Features! Group, There's another new version of APRSISCE for Windows Mobile just posted to the Files section. I discovered a fairly serious resource consumption that causes Windows Mobile to run out of resources and lock up if you drive around at a low zoom level causing lots of OpenStreetMap files to paint. This resource leak has been fixed along with the following changes. 1) Reworked the APRS-IS network handling to improve responsiveness and reduce load on the device. 2) Double-clicking the "APRS OK" window no longer brings up a series of network devices, but instead shows traffic statistics for the APRS-IS connection including packets and bytes sent and received for this session. 3) Reworked where stations are painted on the OSM tiles. They weren't quite landing on streets at larger zoom levels. It uses the old way until all tiles have been filled in, so don't be alarmed if you see the stations "jump" right after the map completes. You should only see this the first time you're in an area while the OSM tiles are loading from the network instead of from the cache. 4) The OSM map is now clipped to the old circle and doesn't overflow around and behind the various other windows. 5) Double-clicking an empty space on the map will popup OpenStreetMap tile statistics if any tiles have been loaded in that session. Eventually, I hope to also show the status of the on-device cache of tiles by zoom level. 6) Increased the available zoom levels. On my 320x240 AT&T Tilt, I can now zoom out to 524 miles instead of the old 256. 7) Reworked how tiles are cached on CE. You might want to locate the OSMTiles folder (probably in the root of your device) and delete it and all files within it. The new caching layer uses directories that exactly mirror OSM's URLs. 8) To store the OSM cache on a Storage Card (assuming you have one), close APRSISCE and edit the APRSISCE.xml file that should have synced to your desktop. Locate the <OSM.Path> argument and set it to <OSM.Path>/Storage Card/OSMTiles/</OSM.Path>. Save the file and wait for it to sync back to your device before running APRSISCE. You should then see an OSMTiles folder on your Storage Card. 9) The OSM Tile Cache now prefetches tiles in anticipation of your movement or zooming. It fetchs one tile out in each ordinal direction from the screen along with all tiles one closer zoom below and one further zoom above. 10) While APRSISCE is fetching OSM tiles, you'll see a yellow circle expanding from the center of the map. Each jump is one tile. When it reaches the outer circle, it has worked off all queued prefetches. It prioritizes the queue, so it'll still get the missing visible tiles first. Let me know what you think and let me know what you'd like to see enhanced and/or changed, especially if you encounter any bugs! Lynn (D) - Author of APRSISCE - (APRS message me at KJ4ERJ-12) ------------------------------------ Yahoo! Groups Links |
|
Lynn Deffenbaugh
Unfortunately, unless you know a good graphics artist that's willing
toggle quoted message
Show quoted text
(and able, I'm not) to draw the standard APRS icons at 32x32 pixels, they'll almost be as hard to see if I stretch them bigger. I'm still working on an algorithm for sizing the icons based on the size of the circle and the zoom level, but I haven't quite got it working right yet. I've got a non-GPS version running on my Vista laptop and I agree, for a large screen (1920x1200 pixels), I agree, they're awful small. I'll see if I can get the algorithm worked out and let you know when it's available. Lynn (D) - KJ4ERJ Noel Shrum wrote:
|
|
Noel Shrum
I have a graphics tablet and The GIMP.
toggle quoted message
Show quoted text
What format are you using for the icons? Perhaps you could send me a copy and I could start working on a larger set for you? Noel Lynn W. Deffenbaugh wrote: Unfortunately, unless you know a good graphics artist that's willing |
|
Lynn Deffenbaugh
Noel,
--- In aprsisce@..., Noel Shrum <noel.w8tvi@...> wrote: I just posted APRSISZoom.zip that contains a zoomed icons for screens larger than 320 pixels. As your scale gets smaller than 2 miles, the icons will begin to expand (the actual scale code is listed below for C programmers). See if they look ok on your higher resolution screen before diving into the editor. But, just to cover all bases, I've also included the .BMP files from which the icons are drawn. The icons are clipped out of these composite pictures automatically by the program. My intention is to allow users to specify external Icon files as well, if they ever want to customize them. For that reason, I'd really rather not build the support for two sizes of icons unless it is visually necessary. Of course, if you do take the time to expand these BMPs and clean them up manually, I could always make the larger set the standard and scale them down on smaller screens! The program assumes that the icons are laid out in a 16x6 grid. It divides the image width and height by 16 and 6 to get the actual icon size. It then assumes a 2 pixel border all around. In the end, it gets a 17x17 pixel icon out of a 21x21 repeating square. Let me know how the expanded icons look on your screen. They didn't actually look as bad as I thought they might when I previewed them here (not on a phone screen, though!) Lynn (D) - KJ4ERJ PS. Here's the current brute-force code that does icon scaling. I want to make it a more dynamic scale based on circle size, device resolution, and the scale of the screen, but, as I mentioned before, I haven't worked out the details yet. It's striking a balance between recognizability and showing lots of stations in a small area. Of course, if you click on a station icon, the popup does tell you what the icon is supposed to be! if (scale >= 128) { Mult = 1; Div = 2; } else if (scale >= 64) { Mult = 2; Div = 3; } else if (scale >= 32) { Mult = 3; Div = 4; } else if (width < 320) /* Drop out all 320x240 screens */ { Mult = 1; Div = 1; } else if (scale >= 8) /* Bigger screens begin to magnify */ { Mult = 1; Div = 1; } else if (scale >= 2) { Mult = 5; Div = 4; } else if (scale >= 0.5) { Mult = 6; Div = 4; } else if (scale >= 0.125) { Mult = 7; Div = 4; } else { Mult = 2; Div = 1; } |
|
Lynn Deffenbaugh
Oh, one more thing. If you do tackle making a larger icon set, it is important that all background pixels be pure white (RGB(255,255,255)). The program uses this as the transparency color when copying the icons to the screen.
toggle quoted message
Show quoted text
Lynn (D) - KJ4ERJ --- In aprsisce@..., "ldeffenb" <kj4erj@...> wrote:
|
|
Noel Shrum
So I was just comparing the two versions of the program,
toggle quoted message
Show quoted text
and I think I figured out what it is I didn't like about the normal one: The icons get too small when you zoom out to see more map. Just too small on the 3" VGA screen for my taste. I do kinda like the bigger icon at higher zoom levels. I don't think they really need to made any bigger.. at least not on this screen. But it would make them look nicer at high zoom, so I'll work on that. Also, what overall size (X/Y) should I make the grid for the larger icon set? Noel W8TVI ldeffenb wrote: Oh, one more thing. If you do tackle making a larger icon set, it is |
|
David Shrum
One feature I would like and that would be to be able to change the
date and time format to the way I want it. Dave N8PU |
|
Lynn Deffenbaugh
Noel,
So I was just comparing the two versions of the program,So, they get too small too quickly in both versions, right? I only expanded them in the second, but didn't change when/how they get small. So the 2x maximum expansion looks good to you? I concur with that opinion as well. Also, what overall size (X/Y) should I make the grid for the larger iconTo get them up to double size clearly, we'll need each square to be 42 pixels instead of the current 21 pixel squares (21-4 giving the 17x17 icons that the program ends up painting). I did a quick experiment here with IrfanView and I think if you start with just a resize/resample 200% retaining aspect ratio, you'll have a good starting point to clean them up. As I mentioned before though, try to make sure that all non-important pixels are actually pure white (RGB(255,255,255)) or they won't go transparent correctly. I'm trying to figure out how to fade them out and keep the transparency so I quit getting the square around fading stations. If you look close at the red dot in the screen caps you sent me, you can see what I'm talking about. I was drawing it at 50% fade, but that messes up the transparency. I'm going back to 100% for the red dot, but aging stations will still start having a square around them as they fade. I'll get it fixed somehow! Thanks again for any efforts you can do on getting clearer, bigger icons. I'm hoping that I can just use them for all resolutions with the 320x240 phone just using them at 50% with Win32 doing the scaling for me. Scaling down tends to look better than scaling up! Lynn (D) - KJ4ERJ |
|
Lynn Deffenbaugh
Dave,
toggle quoted message
Show quoted text
Would a strftime format string be suitable for your time/date format specification? Or how did you have in mind to configure it? http://www.manpagez.com/man/3/strftime/ Lynn (D) - KJ4ERJ David Shrum wrote:
|
|
David Shrum
That page is way above my understanding, but what I had in mind was to be
toggle quoted message
Show quoted text
able to change the way it is displayed so if I wanted to have the month/day/year I could, but if another user wanted it day/month/year or year/day/month they could. On the time it would be nice if we could select a time zone independent of the device, say GMT time for instance we could and be able to select either a 12 or 24 hour format. Since you ask, I thought I would ask. :) I can't remember if you said you were going to try to let us track a specific call, but that would be nice; similar to what it is with http://aprs.fi/?call=n8pu-12 where you can 'type' in a call and if it is available it will display it and scroll the screen so we can view their progress. I know it is much better on a computer screen but if you are out away from yours it would be nice to see where they were. Thanks Dave N8PU -----Original Message-----
From: aprsisce@... [mailto:aprsisce@...] On Behalf Of Lynn W. Deffenbaugh Sent: Monday, August 10, 2009 12:41 PM To: aprsisce@... Subject: Re: [aprsisce] Re: APRSISCE Lock Fix & New Features! Dave, Would a strftime format string be suitable for your time/date format specification? Or how did you have in mind to configure it? http://www.manpagez.com/man/3/strftime/ Lynn (D) - KJ4ERJ David Shrum wrote:
|
|
Lynn Deffenbaugh
--- In aprsisce@..., "David Shrum" <dave.n8pu@...> wrote:
I'll make it much simpler than that page, but it basically says that there'll be %X parameter that represent the various components of the date/time. Whatever you put in the configuration parameter will govern what displays. On the time it would be nice if we could select a time zone independent ofGMT vs the device is not too difficult. Arbitrary time zones are MUCH harder! 12 vs 24 will be automatic given the %X parameters mentioned above. Since you ask, I thought I would ask. :)I'm always willing to listen to any and all ideas. The best part about being the developer is I get to choose what does and does not make the cut (although with me, it's normally WHEN it gets implemented, not IF). I can't remember if you said you were going to try to let us track aYes, that is absolutely in the plans, as is the ability to pan around the map with some way of getting it back to you (or the currently selected station) in the center. Not sure when I'll get that going, but I just added the course/speed parsing. I've got a few more things to do before it's worth putting another release out though. Lynn (D) - KJ4ERJ |
|
Noel Shrum
Lynn W. Deffenbaugh wrote:
Noel,Yeah... they get too small.So I was just comparing the two versions of the program,So, they get too small too quickly in both versions, right? I only I'd rather them just overlap than get really small. I think 2x the "normal" size would be good on a VGA device.I do kinda like the bigger icon at higher zoom levels.So the 2x maximum expansion looks good to you? I concur with that Will do.Also, what overall size (X/Y) should I make the grid for the larger iconTo get them up to double size clearly, we'll need each square to be 42 I'm trying to figure out how to fade them out and keep the transparencyI guess I didn't mess around with that too much... it didn't bother me. Thanks again for any efforts you can do on getting clearer, bigger Noel W8TVI |
|